Mumbai: Visa is shifting from periodic cybersecurity checks to steady, AI-driven defence as cyber threats more and more function at machine pace, with India seeing a pointy rise in cyber fraud.Visa mentioned it has moved from schedule-based vulnerability administration to steady detection, testing, remediation and validation. Its open-source Visa Vulnerability Agentic Harness (VVAH) makes use of a number of AI brokers to find vulnerabilities, check them, develop fixes and validate these fixes earlier than manufacturing. Visa additionally makes use of Imply Time to Adapt (MTTA), relatively than detection or patch counts, to measure how rapidly it may possibly shut exploitable paths.The strategy was examined by means of Anthropic’s Undertaking Glasswing, the place frontier AI fashions have been used in opposition to Visa’s internet-facing and important infrastructure. Visa mentioned its zero-trust structure and community segmentation prevented recognized vulnerabilities from changing into profitable assault chains.“As AI turns into extra highly effective, the cybersecurity problem is now not simply discovering vulnerabilities, it’s fixing them sooner than they are often exploited. India’s digital funds ecosystem has demonstrated what innovation at inhabitants scale can obtain, however sustaining belief would require the identical tempo of innovation in safety and resilience,” mentioned Suresh Sethi, group nation supervisor, Visa in the course of the launch of a whitepaper on Undertaking Glasswing in the course of the World Fintech Fest in Mumbai.Undertaking Glasswing is a defensive cybersecurity initiative during which Visa labored with Anthropic to check frontier AI fashions in opposition to its internet-facing companies, core platforms and important infrastructure. The train uncovered complicated vulnerabilities but additionally confirmed that Visa’s zero-trust structure, community segmentation and different safeguards may break the assault chain earlier than vulnerabilities have been efficiently exploited.Visa has additionally developed an agentic management airplane governing AI brokers’ identities, permissions and working boundaries, whereas its Cost Threats Lab assessments fraud and AI assaults in opposition to fee guidelines and configurations.
Share your ideas within the feedback
Be respectful · TOI community guidelines
It’s working with IBM and Crimson Hat on Undertaking Lightwell to safe open-source software program and has tightened provider necessities round steady vulnerability testing and software program payments of supplies.The urgency is especially excessive in India. Cyber fraud circumstances rose 24% year-on-year to 2.81 million in 2025, whereas losses from funding scams reached Rs 22,495 crore and digital-arrest scams accounted for 9% of losses. The 1930 cyber-fraud helpline logged 32.4 million calls in the course of the 12 months, roughly one each second.